In recent years, with the proliferation of household electrical appliances with network connecting functions, systems for controlling household electrical appliances via a network have been proposed. In addition, with improvements in recognition rate in voice recognition technology, it is becoming possible to input a voiced operation instruction for a household electrical appliance to control the household electrical appliance.
As conventional voice recognition systems, techniques are proposed in which voice information input at a client is transmitted to a server and the server performs voice recognition in order to reduce load on the client with limited processing capability (for example, refer to Patent Literature 1).
In Patent Literature 1 described above, an amount of data transmission between the client and the server is reduced by having the client compression-encode a feature parameter of voice information and transmit the compression-encoded feature parameter.
However, with Patent Literature 1 described above, since the server performs all voice recognition, voice information of all voice recognition objects is transmitted from the client to the server. In other words, all voice information input by the client is transferred to the server. As a result, there is a need to further reduce the amount of data transmission between the client and the server in order to avoid data from concentrating at the server.